forked from mtusi2026/mtusi2026lab7
875 lines
33 KiB
YAML
875 lines
33 KiB
YAML
# Лабораторная 7: стенд lab6 + TLS/mTLS на всех связях, OpenBao, минимальные привилегии.
|
||
#
|
||
# клиент --TLS--> lb :8443 --mTLS (lb)--> api x N (:8443, allowlist identity lb + CRL)
|
||
# api/worker --TLS verify-full--> postgres (роли messenger_api / messenger_worker)
|
||
# api --mTLS + ACL--> redis (кэш сессий)
|
||
# api/worker --mTLS--> broker-gate (allowlist api|worker + CRL) --loopback--> broker (ACL producer/consumer)
|
||
# api/worker --TLS (сквозной, через s3-gate passthrough)--> s3 (раздельные identities по префиксам)
|
||
# api/worker --mTLS--> otel-gate (allowlist api|worker) --loopback--> otel-collector --mTLS--> loki/tempo
|
||
# prometheus --mTLS--> otel-gate (allowlist prometheus); grafana --mTLS--> prometheus/loki/tempo
|
||
# bao-agent-{api,worker} --TLS + AppRole--> openbao (не dev, file storage, ручной unseal)
|
||
# -> tmpfs volume {api,worker}-secrets -> /run/bao (ro) в api/worker
|
||
#
|
||
# Сети (все internal, кроме edge и ops): egress в интернет нет ни у одного сервиса приложения.
|
||
# edge lb (публикует 127.0.0.1:8443) ingress lb, api
|
||
# backend api, worker, postgres, redis, migrate storage api, worker, s3, s3-init
|
||
# queue api, worker, broker(+gate) telemetry-ingest api, worker, otel-collector(+gate)
|
||
# telemetry otel-collector, prometheus, loki, tempo, grafana ops grafana (127.0.0.1:3000)
|
||
# secrets openbao, bao-agent-api, bao-agent-worker (подсеть BAO_SECRETS_SUBNET = secret_id_bound_cidrs)
|
||
# Наружу (только 127.0.0.1): lb 8443 (HTTPS), grafana 3000 (HTTPS + пароль admin).
|
||
name: messenger-lab7
|
||
|
||
# ---------------------------------------------------------------- общие фрагменты
|
||
x-hardened: &hardened
|
||
read_only: true
|
||
cap_drop: [ALL]
|
||
security_opt: ["no-new-privileges:true"]
|
||
|
||
x-queue-env: &queue-env
|
||
BROKER_URL_FILE: /run/bao/broker_url
|
||
BROKER_STREAM: ${BROKER_STREAM:-attachments.jobs}
|
||
BROKER_GROUP: ${BROKER_GROUP:-attachment-workers}
|
||
BROKER_TIMEOUT_SECONDS: ${BROKER_TIMEOUT_SECONDS:-2}
|
||
PIPELINE_VERSION: ${PIPELINE_VERSION:-1}
|
||
JOB_MAX_ATTEMPTS: ${JOB_MAX_ATTEMPTS:-5}
|
||
JOB_TIMEOUT_SECONDS: ${JOB_TIMEOUT_SECONDS:-60}
|
||
JOB_LEASE_GRACE_SECONDS: ${JOB_LEASE_GRACE_SECONDS:-10}
|
||
JOB_CLAIM_IDLE_SECONDS: ${JOB_CLAIM_IDLE_SECONDS:-75}
|
||
JOB_MAX_DELIVERIES: ${JOB_MAX_DELIVERIES:-10}
|
||
JOB_RETRY_BASE_SECONDS: ${JOB_RETRY_BASE_SECONDS:-2}
|
||
JOB_RETRY_MAX_SECONDS: ${JOB_RETRY_MAX_SECONDS:-60}
|
||
STAGING_RETENTION_SECONDS: ${STAGING_RETENTION_SECONDS:-86400}
|
||
RELAY_POLL_INTERVAL_SECONDS: ${RELAY_POLL_INTERVAL_SECONDS:-1}
|
||
RELAY_FAST_PATH_GRACE_SECONDS: ${RELAY_FAST_PATH_GRACE_SECONDS:-2}
|
||
|
||
x-otel-env: &otel-env
|
||
OTEL_SDK_DISABLED: ${OTEL_SDK_DISABLED:-false}
|
||
OTEL_EXPORTER_OTLP_ENDPOINT: https://otel-collector:4318
|
||
OTEL_EXPORTER_OTLP_PROTOCOL: http/protobuf
|
||
OTEL_EXPORTER_OTLP_CERTIFICATE: /run/secrets/ca_crt
|
||
# OTel SDK >= 1.4x задаёт CA на requests.Session, а requests предпочитает REQUESTS_CA_BUNDLE:
|
||
# фиксируем его на CA стенда, чтобы внешняя переменная не подменила доверие экспортёра.
|
||
REQUESTS_CA_BUNDLE: /run/secrets/ca_crt
|
||
OTEL_EXPORTER_OTLP_TIMEOUT: "5"
|
||
OTEL_RESOURCE_ATTRIBUTES: ${OTEL_RESOURCE_ATTRIBUTES:-deployment.environment.name=lab}
|
||
OTEL_SEMCONV_STABILITY_OPT_IN: http
|
||
OTEL_BSP_MAX_QUEUE_SIZE: "2048"
|
||
OTEL_BSP_MAX_EXPORT_BATCH_SIZE: "512"
|
||
OTEL_BSP_EXPORT_TIMEOUT: "5000"
|
||
OTEL_BLRP_MAX_QUEUE_SIZE: "2048"
|
||
OTEL_BLRP_EXPORT_TIMEOUT: "5000"
|
||
OTEL_METRIC_EXPORT_INTERVAL: "10000"
|
||
|
||
x-s3-env: &s3-env
|
||
S3_ENDPOINT: https://s3:8333
|
||
S3_REGION: ${S3_REGION:-us-east-1}
|
||
S3_BUCKET: ${S3_BUCKET:-messenger-attachments}
|
||
S3_ACCESS_KEY_ID_FILE: /run/bao/s3_access_key_id
|
||
S3_SECRET_ACCESS_KEY_FILE: /run/bao/s3_secret_access_key
|
||
S3_CONNECT_TIMEOUT_SECONDS: ${S3_CONNECT_TIMEOUT_SECONDS:-2}
|
||
S3_READ_TIMEOUT_SECONDS: ${S3_READ_TIMEOUT_SECONDS:-10}
|
||
S3_MAX_ATTEMPTS: ${S3_MAX_ATTEMPTS:-2}
|
||
S3_CIRCUIT_OPEN_SECONDS: ${S3_CIRCUIT_OPEN_SECONDS:-5}
|
||
MAX_UPLOAD_BYTES: ${MAX_UPLOAD_BYTES:-10485760}
|
||
MAX_IMAGE_PIXELS: ${MAX_IMAGE_PIXELS:-40000000}
|
||
ATTACHMENT_GC_INTERVAL_SECONDS: ${ATTACHMENT_GC_INTERVAL_SECONDS:-300}
|
||
ATTACHMENT_GC_GRACE_SECONDS: ${ATTACHMENT_GC_GRACE_SECONDS:-900}
|
||
ORPHAN_SWEEP_INTERVAL_SECONDS: ${ORPHAN_SWEEP_INTERVAL_SECONDS:-3600}
|
||
|
||
x-api-env: &api-env
|
||
APP_ENV: ${APP_ENV:-lab}
|
||
APP_VERSION: ${APP_VERSION:-lab7}
|
||
APP_PORT: "8443"
|
||
HEALTH_PORT: "8081"
|
||
OTEL_SERVICE_NAME: messenger-api
|
||
<<: [*queue-env, *otel-env, *s3-env]
|
||
OTEL_EXPORTER_OTLP_CLIENT_CERTIFICATE: /run/secrets/api_crt
|
||
OTEL_EXPORTER_OTLP_CLIENT_KEY: /run/secrets/api_key
|
||
# lab7: HTTPS-сервер с обязательным клиентским сертификатом; allowlist и CRL.
|
||
TLS_CERT_FILE: /run/secrets/api_crt
|
||
TLS_KEY_FILE: /run/secrets/api_key
|
||
TLS_CA_FILE: /run/secrets/ca_crt
|
||
TLS_CLIENT_CERT_FILE: /run/secrets/api_crt
|
||
TLS_CLIENT_KEY_FILE: /run/secrets/api_key
|
||
MTLS_ALLOWED_CLIENTS: lb
|
||
MTLS_CRL_FILE: /run/secrets/crl_pem
|
||
# Секреты — файлы, отрисованные OpenBao Agent (tmpfs volume api-secrets).
|
||
DATABASE_URL_FILE: /run/bao/database_url
|
||
CACHE_URL_FILE: /run/bao/cache_url
|
||
SESSION_STORE: postgres
|
||
SESSION_TTL_SECONDS: ${SESSION_TTL_SECONDS:-3600}
|
||
SESSION_CACHE_TTL_SECONDS: ${SESSION_CACHE_TTL_SECONDS:-60}
|
||
DB_POOL_MIN: ${DB_POOL_MIN:-1}
|
||
DB_POOL_MAX: ${DB_POOL_MAX:-10}
|
||
DB_POOL_TIMEOUT_SECONDS: ${DB_POOL_TIMEOUT_SECONDS:-3}
|
||
DB_POOL_MAX_LIFETIME_SECONDS: ${DB_POOL_MAX_LIFETIME_SECONDS:-600}
|
||
DB_CONNECT_TIMEOUT_SECONDS: ${DB_CONNECT_TIMEOUT_SECONDS:-3}
|
||
DB_STATEMENT_TIMEOUT_MS: ${DB_STATEMENT_TIMEOUT_MS:-5000}
|
||
CACHE_TIMEOUT_SECONDS: ${CACHE_TIMEOUT_SECONDS:-0.25}
|
||
CACHE_RETRY_AFTER_SECONDS: ${CACHE_RETRY_AFTER_SECONDS:-5}
|
||
IDEMPOTENCY_TTL_SECONDS: ${IDEMPOTENCY_TTL_SECONDS:-86400}
|
||
READY_DB_TIMEOUT_SECONDS: ${READY_DB_TIMEOUT_SECONDS:-1}
|
||
DRAIN_SECONDS: ${DRAIN_SECONDS:-3}
|
||
GRACEFUL_TIMEOUT_SECONDS: ${GRACEFUL_TIMEOUT_SECONDS:-10}
|
||
KEEP_ALIVE_SECONDS: ${KEEP_ALIVE_SECONDS:-30}
|
||
# lab7: Argon2id (базовая линия OWASP), бюджет KDF и общий rate limit входа.
|
||
ARGON2_MEMORY_KIB: ${ARGON2_MEMORY_KIB:-19456}
|
||
ARGON2_ITERATIONS: ${ARGON2_ITERATIONS:-2}
|
||
ARGON2_PARALLELISM: ${ARGON2_PARALLELISM:-1}
|
||
PASSWORD_HASH_CONCURRENCY: ${PASSWORD_HASH_CONCURRENCY:-4}
|
||
KDF_QUEUE_TIMEOUT_SECONDS: ${KDF_QUEUE_TIMEOUT_SECONDS:-2}
|
||
LOGIN_RATE_WINDOW_SECONDS: ${LOGIN_RATE_WINDOW_SECONDS:-60}
|
||
LOGIN_RATE_PER_USER: ${LOGIN_RATE_PER_USER:-10}
|
||
LOGIN_RATE_PER_IP: ${LOGIN_RATE_PER_IP:-300}
|
||
MAX_JSON_BODY_BYTES: ${MAX_JSON_BODY_BYTES:-65536}
|
||
RUN_MIGRATIONS_ON_START: "false"
|
||
ORPHAN_SWEEP_ENABLED: "false"
|
||
S3_DOWNLOAD_DEADLINE_SECONDS: ${S3_DOWNLOAD_DEADLINE_SECONDS:-5}
|
||
S3_UPLOAD_DEADLINE_SECONDS: ${S3_UPLOAD_DEADLINE_SECONDS:-10}
|
||
MAX_CONCURRENT_UPLOADS: ${MAX_CONCURRENT_UPLOADS:-4}
|
||
UPLOAD_QUEUE_TIMEOUT_SECONDS: ${UPLOAD_QUEUE_TIMEOUT_SECONDS:-5}
|
||
UPLOAD_READ_TIMEOUT_SECONDS: ${UPLOAD_READ_TIMEOUT_SECONDS:-60}
|
||
IMAGE_DECODE_CONCURRENCY: ${IMAGE_DECODE_CONCURRENCY:-2}
|
||
|
||
x-api-healthcheck: &api-healthcheck
|
||
# Loopback health listener (127.0.0.1:8081, только /health/*); LB проверяет по mTLS.
|
||
test: ["CMD", "python", "-c", "import urllib.request; urllib.request.urlopen('http://127.0.0.1:8081/health/ready', timeout=2).read()"]
|
||
interval: 3s
|
||
timeout: 2s
|
||
retries: 2
|
||
start_period: 20s
|
||
|
||
x-api-limits: &api-limits
|
||
limits:
|
||
cpus: "0.50"
|
||
memory: 320M
|
||
pids: 128
|
||
|
||
x-api-secrets: &api-secrets
|
||
- ca_crt
|
||
- crl_pem
|
||
- api_crt
|
||
- api_key
|
||
|
||
x-api-common: &api-common
|
||
<<: *hardened
|
||
image: messenger-api:${APP_VERSION:-lab7}
|
||
environment: *api-env
|
||
secrets: *api-secrets
|
||
volumes:
|
||
- api-secrets:/run/bao:ro
|
||
tmpfs:
|
||
- /tmp:size=16m,uid=10001,gid=10001
|
||
networks: [ingress, backend, storage, queue, telemetry-ingest]
|
||
dns_opt: ["timeout:1", "attempts:1"]
|
||
|
||
services:
|
||
# ------------------------------------------------------------------ PostgreSQL
|
||
postgres:
|
||
image: postgres:16.4-alpine
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
# Сразу uid 70 (postgres): entrypoint не нужен root/gosu, capabilities не нужны.
|
||
user: "70:70"
|
||
entrypoint: ["/bin/sh", "/opt/messenger/entrypoint.sh"]
|
||
command:
|
||
- postgres
|
||
- -c
|
||
- max_connections=100
|
||
- -c
|
||
- ssl=on
|
||
- -c
|
||
- ssl_cert_file=/run/pgtls/server.crt
|
||
- -c
|
||
- ssl_key_file=/run/pgtls/server.key
|
||
- -c
|
||
- ssl_min_protocol_version=TLSv1.2
|
||
- -c
|
||
- hba_file=/opt/messenger/pg_hba.conf
|
||
- -c
|
||
- password_encryption=scram-sha-256
|
||
environment:
|
||
POSTGRES_USER: messenger
|
||
POSTGRES_DB: messenger
|
||
POSTGRES_PASSWORD_FILE: /run/secrets/postgres_password
|
||
secrets: [postgres_password, postgres_crt, postgres_key, db_api_password, db_worker_password]
|
||
volumes:
|
||
- pgdata:/var/lib/postgresql/data
|
||
- ./infra/postgres/entrypoint.sh:/opt/messenger/entrypoint.sh:ro
|
||
- ./infra/postgres/pg_hba.conf:/opt/messenger/pg_hba.conf:ro
|
||
- ./infra/postgres/initdb:/docker-entrypoint-initdb.d:ro
|
||
tmpfs:
|
||
- /run/pgtls:size=1m,uid=70,gid=70,mode=0700
|
||
- /var/run/postgresql:size=1m,uid=70,gid=70
|
||
- /tmp:size=16m,uid=70,gid=70
|
||
healthcheck:
|
||
test: ["CMD-SHELL", "pg_isready -U messenger -d messenger -h /var/run/postgresql"]
|
||
interval: 3s
|
||
timeout: 3s
|
||
retries: 20
|
||
start_period: 10s
|
||
networks: [backend]
|
||
stop_grace_period: 30s
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "1.00", memory: 512M, pids: 200}
|
||
|
||
# ------------------------------------------------------------------ Redis: кэш сессий
|
||
redis:
|
||
image: redis:7.4.1-alpine
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
user: redis
|
||
entrypoint: ["/bin/sh", "/opt/messenger/cache.sh"]
|
||
secrets: [ca_crt, redis_crt, redis_key, cache_api_password]
|
||
volumes:
|
||
- ./infra/redis/cache.sh:/opt/messenger/cache.sh:ro
|
||
tmpfs: ["/tmp:size=1m", "/data:size=1m"]
|
||
healthcheck:
|
||
test: ["CMD-SHELL", "redis-cli --tls --cacert /run/secrets/ca_crt --cert /run/secrets/redis_crt --key /run/secrets/redis_key -h 127.0.0.1 --user healthcheck --pass x --no-auth-warning ping | grep -q PONG"]
|
||
interval: 3s
|
||
timeout: 3s
|
||
retries: 20
|
||
networks: [backend]
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.25", memory: 96M, pids: 64}
|
||
|
||
# ------------------------------------------------------------------ broker + mTLS gate
|
||
broker:
|
||
image: redis:7.4.1-alpine
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
user: redis
|
||
entrypoint: ["/bin/sh", "/opt/messenger/broker.sh"]
|
||
secrets: [broker_api_password, broker_worker_password]
|
||
volumes:
|
||
- brokerdata:/data
|
||
- ./infra/redis/broker.sh:/opt/messenger/broker.sh:ro
|
||
tmpfs: ["/tmp:size=1m"]
|
||
healthcheck:
|
||
test: ["CMD-SHELL", "redis-cli -h 127.0.0.1 --user healthcheck --pass x --no-auth-warning ping | grep -q PONG"]
|
||
interval: 3s
|
||
timeout: 3s
|
||
retries: 20
|
||
networks: [queue]
|
||
stop_grace_period: 10s
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.50", memory: 320M, pids: 64}
|
||
|
||
broker-gate:
|
||
image: haproxy:2.8.30-alpine
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
# Тот же сетевой namespace, что у broker: Redis слушает только его 127.0.0.1:6379.
|
||
network_mode: "service:broker"
|
||
volumes:
|
||
- ./infra/haproxy/broker-gate.cfg:/usr/local/etc/haproxy/haproxy.cfg:ro
|
||
secrets: [ca_crt, crl_pem, broker_pem]
|
||
tmpfs: ["/tmp:size=1m"]
|
||
depends_on:
|
||
broker:
|
||
condition: service_healthy
|
||
healthcheck:
|
||
test: ["CMD-SHELL", "wget -q -O /dev/null http://127.0.0.1:8405/gate-health"]
|
||
interval: 3s
|
||
timeout: 2s
|
||
retries: 10
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.25", memory: 64M, pids: 64}
|
||
|
||
# ------------------------------------------------------------------ S3 (SeaweedFS)
|
||
s3:
|
||
image: chrislusf/seaweedfs:${SEAWEEDFS_VERSION:-4.50}
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
entrypoint: ["/usr/bin/weed"]
|
||
cap_add: [DAC_OVERRIDE, CHOWN, FOWNER]
|
||
|
||
|
||
command:
|
||
- server
|
||
- -dir=/data
|
||
- -ip=127.0.0.1
|
||
- -ip.bind=127.0.0.1
|
||
- -s3
|
||
# lab7: S3 API и его неаутентифицированный gRPC (8334+10000) — только loopback;
|
||
# в сеть storage порт S3 выдаёт s3-gate (TCP passthrough, TLS сквозной).
|
||
- -s3.port=8334
|
||
- -s3.ip.bind=127.0.0.1
|
||
- -s3.config=/run/secrets/s3_config
|
||
- -s3.iam=false
|
||
- -s3.port.iceberg=0
|
||
- -s3.port.lance=0
|
||
- -s3.autoCreateBucket=false
|
||
# lab7: S3 API только по HTTPS (сертификат s3, SAN DNS:s3).
|
||
- -s3.cert.file=/run/secrets/s3_crt
|
||
- -s3.key.file=/run/secrets/s3_key
|
||
- -s3.concurrentUploadLimitMB=128
|
||
- -master.volumeSizeLimitMB=256
|
||
- -master.electionTimeout=2s
|
||
- -master.telemetry=false
|
||
- -volume.max=0
|
||
environment:
|
||
GOMEMLIMIT: 400MiB
|
||
secrets: [s3_config, s3_crt, s3_key]
|
||
volumes:
|
||
- s3data:/data
|
||
tmpfs: ["/tmp:size=64m"]
|
||
healthcheck:
|
||
# Порт S3 открыт (TLS проверяют s3-init и `ops.py wait-s3` с CA стенда).
|
||
test: ["CMD-SHELL", "nc -z 127.0.0.1 8334 || wget -q -O /dev/null https://127.0.0.1:8334/healthz"]
|
||
interval: 3s
|
||
timeout: 3s
|
||
retries: 20
|
||
start_period: 10s
|
||
networks: [storage]
|
||
stop_grace_period: 10s
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "1.00", memory: 512M, pids: 256}
|
||
|
||
s3-gate:
|
||
image: haproxy:2.8.30-alpine
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
network_mode: "service:s3"
|
||
volumes:
|
||
- ./infra/haproxy/s3-gate.cfg:/usr/local/etc/haproxy/haproxy.cfg:ro
|
||
tmpfs: ["/tmp:size=1m"]
|
||
depends_on:
|
||
s3:
|
||
condition: service_healthy
|
||
healthcheck:
|
||
test: ["CMD-SHELL", "wget -q -O /dev/null http://127.0.0.1:8405/gate-health"]
|
||
interval: 3s
|
||
timeout: 2s
|
||
retries: 10
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.25", memory: 64M, pids: 64}
|
||
|
||
s3-init:
|
||
build: .
|
||
image: messenger-api:${APP_VERSION:-lab7}
|
||
<<: *hardened
|
||
command: ["python", "-m", "src.s3_bootstrap"]
|
||
restart: "no"
|
||
environment:
|
||
S3_ENDPOINT: https://s3:8333
|
||
S3_REGION: ${S3_REGION:-us-east-1}
|
||
S3_BUCKET: ${S3_BUCKET:-messenger-attachments}
|
||
TLS_CA_FILE: /run/secrets/ca_crt
|
||
S3_ADMIN_ACCESS_KEY_ID_FILE: /run/secrets/s3_admin_access_key_id
|
||
S3_ADMIN_SECRET_ACCESS_KEY_FILE: /run/secrets/s3_admin_secret_access_key
|
||
S3_ACCESS_KEY_ID_FILE: /run/secrets/s3_access_key_id
|
||
S3_SECRET_ACCESS_KEY_FILE: /run/secrets/s3_secret_access_key
|
||
S3_WORKER_ACCESS_KEY_ID_FILE: /run/secrets/s3_worker_access_key_id
|
||
S3_WORKER_SECRET_ACCESS_KEY_FILE: /run/secrets/s3_worker_secret_access_key
|
||
secrets: [ca_crt, s3_admin_access_key_id, s3_admin_secret_access_key, s3_access_key_id,
|
||
s3_secret_access_key, s3_worker_access_key_id, s3_worker_secret_access_key]
|
||
tmpfs: ["/tmp:size=8m"]
|
||
depends_on:
|
||
s3-gate:
|
||
condition: service_healthy
|
||
networks: [storage]
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.50", memory: 192M, pids: 64}
|
||
|
||
# ------------------------------------------------------------------ миграции (администратор БД)
|
||
migrate:
|
||
build: .
|
||
image: messenger-api:${APP_VERSION:-lab7}
|
||
<<: *hardened
|
||
command: ["python", "-m", "src.migrate"]
|
||
restart: "no"
|
||
environment:
|
||
DATABASE_URL_FILE: /run/secrets/database_admin_url
|
||
DB_CONNECT_TIMEOUT_SECONDS: ${DB_CONNECT_TIMEOUT_SECONDS:-3}
|
||
DB_STATEMENT_TIMEOUT_MS: "60000"
|
||
MIGRATE_WAIT_SECONDS: "60"
|
||
OTEL_SDK_DISABLED: "true"
|
||
secrets: [database_admin_url, ca_crt]
|
||
tmpfs: ["/tmp:size=8m"]
|
||
depends_on:
|
||
postgres:
|
||
condition: service_healthy
|
||
networks: [backend]
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.50", memory: 256M, pids: 64}
|
||
|
||
# ------------------------------------------------------------------ OpenBao + agents
|
||
openbao:
|
||
image: openbao/openbao:${OPENBAO_VERSION:-2.1.0}
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
user: openbao
|
||
entrypoint: ["/bin/bao"]
|
||
command: ["server", "-config=/openbao/config/server.hcl"]
|
||
secrets: [ca_crt, openbao_crt, openbao_key]
|
||
volumes:
|
||
- baodata:/openbao/file
|
||
- ./infra/openbao/server.hcl:/openbao/config/server.hcl:ro
|
||
tmpfs: ["/tmp:size=8m"]
|
||
healthcheck:
|
||
# 0 = unsealed, 2 = sealed (процесс жив, ждёт оператора), 1 = ошибка.
|
||
test: ["CMD-SHELL", "bao status -address=https://127.0.0.1:8200 -ca-cert=/run/secrets/ca_crt -tls-server-name=openbao >/dev/null 2>&1; test $$? -ne 1"]
|
||
interval: 5s
|
||
timeout: 3s
|
||
retries: 20
|
||
networks: [secrets]
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.50", memory: 256M, pids: 128}
|
||
|
||
bao-agent-api:
|
||
image: openbao/openbao:${OPENBAO_VERSION:-2.1.0}
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
user: "10001:10001"
|
||
entrypoint: ["/bin/bao"]
|
||
command: ["agent", "-config=/openbao/agent/agent.hcl"]
|
||
secrets:
|
||
- ca_crt
|
||
- {source: bao_api_role_id, target: bao_role_id}
|
||
- {source: bao_api_secret_id, target: bao_secret_id}
|
||
volumes:
|
||
- ./infra/openbao/agent-api.hcl:/openbao/agent/agent.hcl:ro
|
||
- api-secrets:/run/bao
|
||
tmpfs: ["/tmp:size=8m"]
|
||
healthcheck:
|
||
test: ["CMD-SHELL", "test -s /run/bao/database_url && test -s /run/bao/cache_url && test -s /run/bao/broker_url && test -s /run/bao/s3_secret_access_key"]
|
||
interval: 3s
|
||
timeout: 2s
|
||
retries: 60
|
||
depends_on:
|
||
openbao:
|
||
condition: service_started
|
||
networks: [secrets]
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.25", memory: 128M, pids: 64}
|
||
|
||
bao-agent-worker:
|
||
image: openbao/openbao:${OPENBAO_VERSION:-2.1.0}
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
user: "10001:10001"
|
||
entrypoint: ["/bin/bao"]
|
||
command: ["agent", "-config=/openbao/agent/agent.hcl"]
|
||
secrets:
|
||
- ca_crt
|
||
- {source: bao_worker_role_id, target: bao_role_id}
|
||
- {source: bao_worker_secret_id, target: bao_secret_id}
|
||
volumes:
|
||
- ./infra/openbao/agent-worker.hcl:/openbao/agent/agent.hcl:ro
|
||
- worker-secrets:/run/bao
|
||
tmpfs: ["/tmp:size=8m"]
|
||
healthcheck:
|
||
test: ["CMD-SHELL", "test -s /run/bao/database_url && test -s /run/bao/broker_url && test -s /run/bao/s3_secret_access_key"]
|
||
interval: 3s
|
||
timeout: 2s
|
||
retries: 60
|
||
depends_on:
|
||
openbao:
|
||
condition: service_started
|
||
networks: [secrets]
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.25", memory: 128M, pids: 64}
|
||
|
||
# ------------------------------------------------------------------ API и worker
|
||
api:
|
||
build: .
|
||
<<: *api-common
|
||
restart: unless-stopped
|
||
depends_on:
|
||
postgres:
|
||
condition: service_healthy
|
||
redis:
|
||
condition: service_healthy
|
||
migrate:
|
||
condition: service_completed_successfully
|
||
s3-init:
|
||
condition: service_completed_successfully
|
||
bao-agent-api:
|
||
condition: service_healthy
|
||
# Только сеть ingress (LB) и зависимости; порт 8443 — mTLS, наружу не публикуется.
|
||
expose: ["8443"]
|
||
healthcheck: *api-healthcheck
|
||
stop_signal: SIGTERM
|
||
stop_grace_period: 20s
|
||
deploy:
|
||
replicas: ${API_REPLICAS:-2}
|
||
resources: *api-limits
|
||
|
||
worker:
|
||
build: .
|
||
image: messenger-api:${APP_VERSION:-lab7}
|
||
<<: *hardened
|
||
command: ["python", "-m", "src.worker"]
|
||
restart: unless-stopped
|
||
environment:
|
||
<<: [*queue-env, *otel-env, *s3-env]
|
||
OTEL_SERVICE_NAME: messenger-worker
|
||
OTEL_EXPORTER_OTLP_CLIENT_CERTIFICATE: /run/secrets/worker_crt
|
||
OTEL_EXPORTER_OTLP_CLIENT_KEY: /run/secrets/worker_key
|
||
APP_VERSION: ${APP_VERSION:-lab7}
|
||
APP_ENV: ${APP_ENV:-lab}
|
||
INSTANCE_ID_MODE: hostname
|
||
TLS_CA_FILE: /run/secrets/ca_crt
|
||
TLS_CLIENT_CERT_FILE: /run/secrets/worker_crt
|
||
TLS_CLIENT_KEY_FILE: /run/secrets/worker_key
|
||
DATABASE_URL_FILE: /run/bao/database_url
|
||
DB_POOL_MIN: "1"
|
||
DB_POOL_MAX: ${WORKER_DB_POOL_MAX:-6}
|
||
DB_POOL_TIMEOUT_SECONDS: ${DB_POOL_TIMEOUT_SECONDS:-3}
|
||
DB_CONNECT_TIMEOUT_SECONDS: ${DB_CONNECT_TIMEOUT_SECONDS:-3}
|
||
DB_STATEMENT_TIMEOUT_MS: ${DB_STATEMENT_TIMEOUT_MS:-5000}
|
||
# Upload-worker: CPU/RAM/время одного изображения ограничены (подпроцесс + RLIMIT_AS + timeout).
|
||
WORKER_CONCURRENCY: ${WORKER_CONCURRENCY:-2}
|
||
WORKER_IMAGE_TIMEOUT_SECONDS: ${WORKER_IMAGE_TIMEOUT_SECONDS:-30}
|
||
WORKER_IMAGE_MEMORY_MB: ${WORKER_IMAGE_MEMORY_MB:-1024}
|
||
WORKER_FAULTS: "false"
|
||
secrets: [ca_crt, worker_crt, worker_key]
|
||
volumes:
|
||
- worker-secrets:/run/bao:ro
|
||
depends_on:
|
||
postgres:
|
||
condition: service_healthy
|
||
broker-gate:
|
||
condition: service_healthy
|
||
migrate:
|
||
condition: service_completed_successfully
|
||
s3-init:
|
||
condition: service_completed_successfully
|
||
bao-agent-worker:
|
||
condition: service_healthy
|
||
healthcheck:
|
||
test: ["CMD", "python", "-c", "import os,time,sys; sys.exit(0 if time.time()-os.path.getmtime('/tmp/worker-heartbeat') < 90 else 1)"]
|
||
interval: 5s
|
||
timeout: 3s
|
||
retries: 3
|
||
start_period: 20s
|
||
stop_signal: SIGTERM
|
||
stop_grace_period: 75s
|
||
tmpfs:
|
||
- /tmp:size=16m,uid=10001,gid=10001
|
||
networks: [backend, storage, queue, telemetry-ingest]
|
||
dns_opt: ["timeout:1", "attempts:1"]
|
||
deploy:
|
||
replicas: ${WORKER_REPLICAS:-1}
|
||
resources:
|
||
limits: {cpus: "1.00", memory: 1280M, pids: 128}
|
||
|
||
# Экземпляр с коротким TTL — только для автотеста истечения сессии (профиль ttl).
|
||
api-short-ttl:
|
||
<<: *api-common
|
||
profiles: [ttl]
|
||
restart: "no"
|
||
environment:
|
||
<<: *api-env
|
||
SESSION_TTL_SECONDS: ${SHORT_SESSION_TTL_SECONDS:-2}
|
||
depends_on:
|
||
api:
|
||
condition: service_healthy
|
||
expose: ["8443"]
|
||
healthcheck: *api-healthcheck
|
||
deploy:
|
||
resources: *api-limits
|
||
|
||
# ------------------------------------------------------------------ LB
|
||
lb:
|
||
image: haproxy:2.8.30-alpine
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
ports:
|
||
- "127.0.0.1:${HTTPS_PORT:-8443}:8443"
|
||
volumes:
|
||
- ./infra/haproxy/haproxy.cfg:/usr/local/etc/haproxy/haproxy.cfg:ro
|
||
secrets: [ca_crt, lb_public_pem, lb_client_pem]
|
||
tmpfs: ["/tmp:size=1m"]
|
||
depends_on:
|
||
api:
|
||
condition: service_healthy
|
||
healthcheck:
|
||
test: ["CMD-SHELL", "wget -q -O /dev/null http://127.0.0.1:8404/lb-health"]
|
||
interval: 5s
|
||
timeout: 2s
|
||
retries: 3
|
||
stop_grace_period: 5s
|
||
networks: [edge, ingress]
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.50", memory: 128M, pids: 64}
|
||
|
||
# ------------------------------------------------------------------ наблюдаемость
|
||
otel-collector:
|
||
image: otel/opentelemetry-collector-contrib:0.115.1
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
command: ["--config=/etc/otelcol/collector.yaml"]
|
||
volumes:
|
||
- ./infra/otel/collector.yaml:/etc/otelcol/collector.yaml:ro
|
||
secrets: [ca_crt, otel_collector_crt, otel_collector_key]
|
||
environment:
|
||
GOMEMLIMIT: 200MiB
|
||
networks: [telemetry-ingest, telemetry]
|
||
stop_grace_period: 10s
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.50", memory: 384M, pids: 128}
|
||
|
||
otel-gate:
|
||
image: haproxy:2.8.30-alpine
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
network_mode: "service:otel-collector"
|
||
volumes:
|
||
- ./infra/haproxy/otel-gate.cfg:/usr/local/etc/haproxy/haproxy.cfg:ro
|
||
secrets: [ca_crt, crl_pem, otel_collector_pem]
|
||
tmpfs: ["/tmp:size=1m"]
|
||
depends_on: [otel-collector]
|
||
healthcheck:
|
||
test: ["CMD-SHELL", "wget -q -O /dev/null http://127.0.0.1:8405/gate-health"]
|
||
interval: 5s
|
||
timeout: 2s
|
||
retries: 10
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.25", memory: 64M, pids: 64}
|
||
|
||
prometheus:
|
||
image: prom/prometheus:v2.55.1
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
command:
|
||
- --config.file=/etc/prometheus/prometheus.yml
|
||
- --web.config.file=/etc/prometheus/web.yml
|
||
- --storage.tsdb.path=/prometheus
|
||
- --storage.tsdb.retention.time=7d
|
||
volumes:
|
||
- ./infra/prometheus/prometheus.yml:/etc/prometheus/prometheus.yml:ro
|
||
- ./infra/prometheus/web.yml:/etc/prometheus/web.yml:ro
|
||
- ./infra/prometheus/client.yml:/etc/prometheus/client.yml:ro
|
||
- ./infra/prometheus/rules:/etc/prometheus/rules:ro
|
||
- promdata:/prometheus
|
||
secrets: [ca_crt, prometheus_crt, prometheus_key]
|
||
healthcheck:
|
||
test: ["CMD", "promtool", "check", "healthy", "--url=https://127.0.0.1:9090", "--http.config.file=/etc/prometheus/client.yml"]
|
||
interval: 5s
|
||
timeout: 3s
|
||
retries: 20
|
||
networks: [telemetry]
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.50", memory: 512M, pids: 128}
|
||
|
||
loki:
|
||
image: grafana/loki:3.3.2
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
command: ["-config.file=/etc/loki/loki.yaml"]
|
||
volumes:
|
||
- ./infra/loki/loki.yaml:/etc/loki/loki.yaml:ro
|
||
- lokidata:/loki
|
||
secrets: [ca_crt, loki_crt, loki_key]
|
||
tmpfs: ["/tmp:size=16m"]
|
||
networks: [telemetry]
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.50", memory: 512M, pids: 128}
|
||
|
||
tempo-init:
|
||
image: grafana/tempo:2.6.1
|
||
user: root
|
||
entrypoint: ["chown", "10001:10001", "/var/tempo"]
|
||
restart: "no"
|
||
cap_drop: [ALL]
|
||
cap_add: [CHOWN]
|
||
security_opt: ["no-new-privileges:true"]
|
||
volumes:
|
||
- tempodata:/var/tempo
|
||
network_mode: none
|
||
|
||
tempo:
|
||
image: grafana/tempo:2.6.1
|
||
<<: *hardened
|
||
restart: unless-stopped
|
||
command: ["-config.file=/etc/tempo/tempo.yaml"]
|
||
volumes:
|
||
- ./infra/tempo/tempo.yaml:/etc/tempo/tempo.yaml:ro
|
||
- tempodata:/var/tempo
|
||
secrets: [ca_crt, tempo_crt, tempo_key]
|
||
tmpfs: ["/tmp:size=16m"]
|
||
depends_on:
|
||
tempo-init:
|
||
condition: service_completed_successfully
|
||
networks: [telemetry]
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.50", memory: 512M, pids: 128}
|
||
|
||
grafana:
|
||
image: grafana/grafana:11.4.0
|
||
cap_drop: [ALL]
|
||
security_opt: ["no-new-privileges:true"]
|
||
restart: unless-stopped
|
||
environment:
|
||
GF_SERVER_PROTOCOL: https
|
||
GF_SERVER_CERT_FILE: /run/secrets/grafana_crt
|
||
GF_SERVER_CERT_KEY: /run/secrets/grafana_key
|
||
GF_SERVER_ROOT_URL: https://localhost:${GRAFANA_PORT:-3000}/
|
||
GF_SECURITY_ADMIN_USER: admin
|
||
GF_SECURITY_ADMIN_PASSWORD__FILE: /run/secrets/grafana_admin_password
|
||
GF_SECURITY_COOKIE_SECURE: "true"
|
||
GF_USERS_ALLOW_SIGN_UP: "false"
|
||
GF_AUTH_ANONYMOUS_ENABLED: "false"
|
||
GF_ANALYTICS_REPORTING_ENABLED: "false"
|
||
GF_ANALYTICS_CHECK_FOR_UPDATES: "false"
|
||
GF_PLUGINS_PLUGIN_ADMIN_ENABLED: "false"
|
||
GF_DASHBOARDS_DEFAULT_HOME_DASHBOARD_PATH: /var/lib/grafana/dashboards/messenger.json
|
||
secrets: [grafana_admin_password, ca_crt, grafana_crt, grafana_key]
|
||
volumes:
|
||
- ./infra/grafana/provisioning:/etc/grafana/provisioning:ro
|
||
- ./infra/grafana/dashboards:/var/lib/grafana/dashboards:ro
|
||
- grafanadata:/var/lib/grafana
|
||
ports:
|
||
- "127.0.0.1:${GRAFANA_PORT:-3000}:3000"
|
||
depends_on:
|
||
prometheus:
|
||
condition: service_healthy
|
||
networks: [telemetry, ops]
|
||
deploy:
|
||
resources:
|
||
limits: {cpus: "0.50", memory: 256M, pids: 128}
|
||
|
||
# ------------------------------------------------------------------ инструменты оператора
|
||
# PKI без Python-зависимостей на хосте: docker compose --profile tools run --rm tools scripts/pki.py init
|
||
tools:
|
||
build: .
|
||
image: messenger-api:${APP_VERSION:-lab7}
|
||
profiles: [tools]
|
||
user: "0:0"
|
||
entrypoint: ["python"]
|
||
working_dir: /work
|
||
network_mode: none
|
||
cap_drop: [ALL]
|
||
cap_add: [CHOWN, DAC_OVERRIDE, FOWNER]
|
||
security_opt: ["no-new-privileges:true"]
|
||
volumes:
|
||
- ./scripts:/work/scripts:ro
|
||
- ./secrets:/work/secrets
|
||
|
||
# TLS/ACL-аудит всей матрицы (student-tests/security/tls_audit.py). Подключается к каждой
|
||
# внутренней сети как «посторонний клиент»; получает только тестовые материалы оператора.
|
||
audit:
|
||
image: messenger-api:${APP_VERSION:-lab7}
|
||
profiles: [audit]
|
||
<<: *hardened
|
||
entrypoint: ["python"]
|
||
working_dir: /app
|
||
environment:
|
||
AUDIT_SECRETS: /run/secrets
|
||
OTEL_SDK_DISABLED: "true"
|
||
secrets: [ca_crt, crl_pem, intruder_crt, intruder_key, lb_client_pem, api_crt, api_key, worker_crt, worker_key,
|
||
prometheus_crt, prometheus_key, grafana_crt, grafana_key, redis_crt, redis_key,
|
||
cache_api_password, broker_api_password, broker_worker_password,
|
||
db_api_password, db_worker_password, s3_access_key_id, s3_secret_access_key,
|
||
s3_worker_access_key_id, s3_worker_secret_access_key]
|
||
volumes:
|
||
- ./student-tests:/app/student-tests:ro
|
||
tmpfs: ["/tmp:size=16m,uid=10001,gid=10001"]
|
||
networks: [ingress, backend, storage, queue, telemetry-ingest, telemetry, secrets]
|
||
|
||
volumes:
|
||
pgdata:
|
||
s3data:
|
||
brokerdata:
|
||
baodata:
|
||
promdata:
|
||
lokidata:
|
||
tempodata:
|
||
grafanadata:
|
||
# Секреты, отрисованные OpenBao Agent: только память (tmpfs), владелец uid 10001,
|
||
# переживают рестарт контейнеров api/worker, но не рестарт Docker (агент отрисует заново).
|
||
api-secrets:
|
||
driver: local
|
||
driver_opts: {type: tmpfs, device: tmpfs, o: "size=1m,uid=10001,gid=10001,mode=0700"}
|
||
worker-secrets:
|
||
driver: local
|
||
driver_opts: {type: tmpfs, device: tmpfs, o: "size=1m,uid=10001,gid=10001,mode=0700"}
|
||
|
||
networks:
|
||
edge: {}
|
||
ops: {}
|
||
ingress: {internal: true}
|
||
backend: {internal: true}
|
||
storage: {internal: true}
|
||
queue: {internal: true}
|
||
telemetry-ingest: {internal: true}
|
||
telemetry: {internal: true}
|
||
secrets:
|
||
internal: true
|
||
ipam:
|
||
config:
|
||
- subnet: ${BAO_SECRETS_SUBNET:-172.30.250.0/24}
|
||
|
||
secrets:
|
||
# PKI (scripts/pki.py): ca.key НЕ монтируется никуда.
|
||
ca_crt: {file: ./secrets/pki/ca.crt}
|
||
crl_pem: {file: ./secrets/pki/crl.pem}
|
||
lb_public_pem: {file: ./secrets/pki/lb-public.pem}
|
||
lb_client_pem: {file: ./secrets/pki/lb-client.pem}
|
||
api_crt: {file: ./secrets/pki/api.crt}
|
||
api_key: {file: ./secrets/pki/api.key}
|
||
worker_crt: {file: ./secrets/pki/worker.crt}
|
||
worker_key: {file: ./secrets/pki/worker.key}
|
||
postgres_crt: {file: ./secrets/pki/postgres.crt}
|
||
postgres_key: {file: ./secrets/pki/postgres.key}
|
||
redis_crt: {file: ./secrets/pki/redis.crt}
|
||
redis_key: {file: ./secrets/pki/redis.key}
|
||
broker_pem: {file: ./secrets/pki/broker.pem}
|
||
s3_crt: {file: ./secrets/pki/s3.crt}
|
||
s3_key: {file: ./secrets/pki/s3.key}
|
||
otel_collector_crt: {file: ./secrets/pki/otel-collector.crt}
|
||
otel_collector_key: {file: ./secrets/pki/otel-collector.key}
|
||
otel_collector_pem: {file: ./secrets/pki/otel-collector.pem}
|
||
loki_crt: {file: ./secrets/pki/loki.crt}
|
||
loki_key: {file: ./secrets/pki/loki.key}
|
||
tempo_crt: {file: ./secrets/pki/tempo.crt}
|
||
tempo_key: {file: ./secrets/pki/tempo.key}
|
||
prometheus_crt: {file: ./secrets/pki/prometheus.crt}
|
||
prometheus_key: {file: ./secrets/pki/prometheus.key}
|
||
grafana_crt: {file: ./secrets/pki/grafana.crt}
|
||
grafana_key: {file: ./secrets/pki/grafana.key}
|
||
openbao_crt: {file: ./secrets/pki/openbao.crt}
|
||
openbao_key: {file: ./secrets/pki/openbao.key}
|
||
intruder_crt: {file: ./secrets/pki/intruder.crt}
|
||
intruder_key: {file: ./secrets/pki/intruder.key}
|
||
# Учётные данные (scripts/init_secrets.py). Приложению — только через OpenBao.
|
||
postgres_password: {file: ./secrets/postgres_password}
|
||
database_admin_url: {file: ./secrets/database_admin_url}
|
||
db_api_password: {file: ./secrets/db_api_password}
|
||
db_worker_password: {file: ./secrets/db_worker_password}
|
||
cache_api_password: {file: ./secrets/cache_api_password}
|
||
broker_api_password: {file: ./secrets/broker_api_password}
|
||
broker_worker_password: {file: ./secrets/broker_worker_password}
|
||
s3_config: {file: ./secrets/s3_config.json}
|
||
s3_admin_access_key_id: {file: ./secrets/s3_admin_access_key_id}
|
||
s3_admin_secret_access_key: {file: ./secrets/s3_admin_secret_access_key}
|
||
s3_access_key_id: {file: ./secrets/s3_access_key_id}
|
||
s3_secret_access_key: {file: ./secrets/s3_secret_access_key}
|
||
s3_worker_access_key_id: {file: ./secrets/s3_worker_access_key_id}
|
||
s3_worker_secret_access_key: {file: ./secrets/s3_worker_secret_access_key}
|
||
grafana_admin_password: {file: ./secrets/grafana_admin_password}
|
||
# AppRole (scripts/bao.py): только для агентов своей роли.
|
||
bao_api_role_id: {file: ./secrets/bao/api_role_id}
|
||
bao_api_secret_id: {file: ./secrets/bao/api_secret_id}
|
||
bao_worker_role_id: {file: ./secrets/bao/worker_role_id}
|
||
bao_worker_secret_id: {file: ./secrets/bao/worker_secret_id}
|